Of all radio telephone sets mounted in vehicles, such as a mobile telephone apparatus, an apparatus has a hands-free speech mode in addition to a handset talking mode. The hands-free speech mode is a mode in which a receiving speech loudspeaker and a transmitting speech microphone are arranged in a telephone apparatus body or a dashboard independently of a handset, and the loudspeaker and the microphone constitute a handset to perform talking. In this hands-free speech mode, a speaker can speak without holding a handset. The speaker can safely speak without having to hold the steering wheel with a single hand even during driving. The hands-free speech mode is very useful to improve safety during driving.
In such a hands-free speech mode, however, a received speech output from the loudspeaker is input to the microphone during talking to produce an acoustic echo. This echo signal is often transmitted to the partner side. Therefore, an echo canceler for canceling the above echo signal is generally arranged in a telephone apparatus having such a hands-free speech mode. In a radio telephone apparatus for encoding a speech signal into digital data and transmitting the digital data, the echo canceler is inevitably used because signal delays caused by a speech encoding/decoding circuit (speech codec) and an error correction encoding/decoding circuit (channel codec) are large so that the acoustic echoes bother the speaker.
A conventional acoustic echo canceler comprises a digital signal processor. This processor comprises arithmetic operation means for simulating acoustic characteristics of an echo path from the loudspeaker to the microphone, subtracting means for obtaining a difference between an output from the arithmetic operation means and an input from the microphone, and means for changing the acoustic characteristics of the arithmetic operation means such that the output from the subtracter becomes zero. In this processor, a speech signal supplied to the loudspeaker is supplied to the arithmetic operation means to generate a pseudo acoustic echo, and this pseudo acoustic echo is subtracted from an actual transmitting speech signal (containing the acoustic echo) transmitted from the microphone, thereby canceling the acoustic echo. An output from the subtracter is deviated from zero in accordance with a change in acoustic characteristics of the echo path. However, when the coefficient of the arithmetic operation means is changed such that the output from the subtracter becomes zero, the acoustic echo signal can always be canceled.
The following problem to be solved is left in a conventional speech communication apparatus having such an echo canceler. A volume control is inserted in the receiving speech signal path of the speech communication apparatus. A speaker operates this volume control to adjust the receiving speech volume. When the receiving speech volume is set excessively high upon operation of the volume control, the signal level of a receiving speech signal exceeds the dynamic ranges of a receiving speech amplifier and a loudspeaker, thereby distorting received speech. When the received speech produced from the loudspeaker is distorted, an echo signal having distortion corresponding to the distortion of the received speech is also input to the microphone. As a result, the echo canceler cannot sufficiently cancel the echo signal.
Even if the volume control is adjusted so as to obtain an optimal receiving speech volume, the receiving speech signal level may be temporarily increased due to a small loss of the wired line of the partner or a change in loudness of the partner's voice. In this case, the received Speech is distorted, and the distorted echo may not be canceled.
The above description have exemplified the hands-free speech mode. However, even in the handset talking mode, when the handset is kept away from a face, speech received from the receiver may be input to the microphone, and an acoustic echo signal may be generated. That is, the above problem is not limited to the hands-free speech mode. In addition, the above problem is not limited to a mobile telephone apparatus, but is equally applicable to a wired telephone apparatus, a radio communication machine, and a transceiver, and the like.
The present invention has been made in consideration of the above situation, and has as its object to cause an echo canceler to sufficiently cancel an acoustic echo even if the signal level of a receiving speech signal exceeds the dynamic ranges of a receiving speech amplifier and a loudspeaker in a speech communication apparatus having the echo canceler for canceling an acoustic echo contained in a transmitted signal by estimating an echo path of the received signal, thereby further improving talking quality.